relationships between geogenic radon potential and gamma ...14th international workshop –garrm,...
TRANSCRIPT
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Relationships between geogenic radon potential and gamma ray maps with indoor radon levels at Caprarola
municipality (central Italy).1Ruggiero L., 1Bigi S., 2Brilli M., 2,3Ciotoli G., 4Dehandschutter B., 3Galli G.,
2Giustini F., 1Lombardi S., 3Lucchetti C., 2Pennica F., 3Pizzino L., 3Sciarra A., 2Sirianni P., 1Tartarello M. C., 2Voltaggio M.
1 Dipartimento di Scienze della Terra, Sapienza-Università di Roma, DST-Sapienza, Italy2 National Research Council, Institute of Environmental Geology and Geoengineering, CNR-IGAG, Italy
3 Istituto Nazionale di Geofisica e Vulcanologia, Italy4 Federal Agency for Nuclear Control - Belgium
14th INTERNATIONAL WORKSHOP - GARRM
GEOLOGICAL ASPECTS OF RADON RISK MAPPING
September 18th – 20th(21nd), 2018, Prague, Czech Republic
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
The LIFE-Respire partnership
Centro di Ricerca, Previsione, Prevenzione e Controllo dei Rischi Geologici,
Sapienza Università di Roma, Rome – Italy (project leader)
Istituto Nazionale di Geofisica e Vulcanologia, Rome - Italy
Istituto di Geologia Ambientale e Geoingegneria,
Consiglio Nazionale delle Ricerche, Rome - Italy
Federal Agency for Nuclear Control, Brussels, Belgium
Elica S.p.A., Fabriano (AN), Italy
Re
se
arc
h In
stitu
tes
Industry
DURATION:
Start date 01/09/2017 End date 31/08/2020
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
The LIFE-Respire project objectives
1. To demonstrate in 4 significant areas, with different GRP in Italy and Belgium, a cost-
effective and eco-friendly solution for Rn real-time measurement and remediation to keep indoor
Rn levels below 300 Bq/m3 level (as indicated in European Directive 2013/59/EURATOM).
2. To realize and install an intelligent and adaptable hybrid Rn remediation system composed by
sensors, an Air Quality Balancer (called SNAP) and an external additional fan-system (eolian and/or
electric) working on positive pressure method.
3. To construct a real time LIFE-RESPIRE geodatabase of collected continuous Rn measurements,
coupled with other geological and geochemical data, as well as building characteristics.
4. To provide local authorities with Rn hazard guidelines, i.e., guidance about Indoor Radon Abatement
Acts, and about the risk, measurement and remediation of radon in workplaces and schools; and real-time
WebGis radon maps for land use planning and health risk assessment, helping to prepare relevant national
action plans (Articles 54, 74 and 103 in 2013/59/EURATOM).
Radon rEal time monitoring System and Proactive Indoor Remediation
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Respire project activities
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
1. To construct the geogenic radon potential (GRP) map of the
Caprarola municipality (northern Lazio, central Italy) by using
multivariate regression model (Empirical Bayesian Kriging
Regression, EBKR)
2. To study the relationships between:
• GRP and Rn indoor
• Terrestrial gamma dose rate and Rn indoor
• Indoor gamma dose rate and Rn indoor.
Objectives
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Rationale
The Geogenic Radon Potential (GRP) defines the availability of radon generated in the ground for surface exhalation or infiltration into buildings.
GRP of an area is directly measurable in the field by the Rn concentrations in the soil pores. These quantities actually are available only regionally and/or locally.
6
Other factors control the GRP of a region: the source and the transport in the ground.
The transport is additionally affected directly/indirectly by factors related to tectonics, and geodynamic features i.e., active faults, seismicity, geothermal activity and volcanism.
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
The indoor radon mapping problem
7
The spatial distribution of indoor Rn samples at large scale is oftenclustered
Indoor Rn usually shows strong variability at least on short geographicscale (i.e., non-autocorrelated and non-stationary spatial behavior)because of its multifactorial dependence (i.e., physical,meteorological and anthropic parameters)
The direct interpolation of indoor Rn values beyond the boundary ofan urban area could be a difficult and non-robust procedure toaccomplish
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
The radon mapping problem
• Alternative approaches consider the costruction of GeogenicRadon Potential (GRP) maps by using available geological andgeochemical information (i.e., rock permeability, faults,radionuclide content, etc.), calibrated by using soil gas radon.
• These data are more suitable to construct GRP maps becausethey are characterised by:
higher spatial autocorrelation
lower variability
not depend by anthropogenic factors
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
MethodsHow Empirical Bayesian Kriging Regression works
Empirical Bayesian Kriging Regression (EBKR) combines ordinary least square
(OLS) regression and simple kriging to provide accurate predictions of
moderately non-stationary data at a local level. It accounts for these local
effects by dividing the input data into subsets of a given size before doing any
modeling.
EBKR estimates the semivariogram through a process of subsetting and
repeated simulations and transforms the proxy variables into their principal
components, that are used as the explanatory variables in the regression
model. The PCs transformation also solves the problem of multicollinearity
because each PC is uncorrelated with the others.
Each principal component captures a certain proportion of the total variability
of the explanatory variables; (in general few principal components up to 70%
of the total variability).
Simulated semivariograms. Empirical semivariances fall in the middle of the spectrum
Prediction with subsets
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Disadvantages of EBKR
Processing is slower than other kriging methods
Anisotropy is unavailable
Advantages of EBKR
Requires minimal interactive modelling
Allows accurate predictions of non-stationary data
Uses local models to capture small scale effects
More accurate than other kriging methods for small datasets
MethodsEmpirical Bayesian Kriging Regression
Hengl T., Heuvelink G.B.M.,Rossiter D.G. (2007). About
regression-kriging: From equations to case studies.
Computers & Geosciences, 33 (10),1301-1315
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
The study areaCaprarola municipality (central Italy)
11
The study area is located about 100 km N of Rome, in central Italy.
Area: 58 km2
Population (2017): 5417
Density: 94/ km2
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
MethodsGeological setting
- volcanic rocks (tuffs and lavas), outcropping in whole area.
- sedimentary rocks (recent and lake deposits, mainly outcropping in the Western sector of the area.
• The volcanic products are particularly enriched on natural radionuclides, and they were traditionally used as building materials in the old center of the Caprarola village.
Simplified geology of the Caprarola municipality modified after
the Geological Map of the Lazio Region (scale 1:25000)
(Cosentino and Pasquali, 2012).Soil samples
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
MethodsThe Proxy variables
EMAN = Emanation coefficient, averaged on lithology →
Diffusive RnFLUX from the soil, averaged on lithology and calculated by →
GAMMA = Terrestrial gamma dose rate → 187 samples
(NaI γ-ray portable scintillometer Scintrex GRS-500) Ordinary Kriging
CO2 = soil-gas CO2 concentration → 178 samples (Draeger X-am 7000) Ordinary Kriging
Perm = Permeability of soil/rocks → (measured in then field at the soil gas sampling points)
𝜂 =𝑅𝑛𝑐𝑝
𝑅𝑛𝑡𝑜𝑡=
𝑅𝑛 𝑎𝑚𝑜𝑢𝑛𝑡 𝑖𝑛 𝑐𝑜𝑛𝑛𝑒𝑐𝑡𝑒𝑑 𝑝𝑜𝑟𝑒𝑠
𝑡𝑜𝑡𝑎𝑙 𝑅𝑛 𝑎𝑚𝑜𝑢𝑛𝑡(Sasaki et al., 2004)
𝐽 = 𝜂𝜌𝑝 1 − 𝜑 226𝑅𝑎 𝜆𝐷0 𝑒− 6𝜑𝑅𝑠+6𝑅𝑠
14𝜑
(Voltaggio et al., 2006)
Radionuclide content in Bq/kg (226Ra, 238U, 232Th, 40K) averaged on lithological types. 16 samples (high-resolution gamma spectrometer equipped with a low-background HPGe coaxial detector, GEM – EG&G ORTEC)
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
ResultsResponse variable: Soil gas radon
Soil gas radon sampling
Soil gas and permeability measurements
Mainstatistics N Mean (95% CI) GM (95% CI) Min Max St.dev
222Rn (kBq/m3) 180 157(143-172) 130 (118-143) 17 865 99.2
220Rn (kBq/m3) 180 314(286-341) 259 (235-286) 34 >1000 183
CO2 (%, v/v) 178 0.68 (0.61-0.65) 0.55 (0.49-0.61) 0.06 3.1 0.47
K (m2) 157 1.19E-11 4.13E-13 1.12E-10 1.4E-11
TGDR (mS/h) 187 0.36 (0.34-0.38) 0.34 (0.33-0.36) 0.15 0.97 0.11
226Ra (Bq/kg) 16 149 (120-179) 79 318 57
238U (Bq/kg) 16 181 (142-219) 85 369 18
232Th (Bq/kg) 16 253 (200-307) 23 481 104
40K (Bq/kg) 16 744 (611-877) 317 1236 249
222Rn Emanation 16 0.087 (0.076-0.099)
0.046 0.14 0.021
222Rn flux (Bq/m2/day)
16 7291(6209-8373)
3730 11482 2030
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Results - Proxy variable maps
Permeability (m2)
Radionuclide content (U, K, Ra, Th)
Emanation coefficient
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
ResultsPermeability – Terrestrial gamma dose rate
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Cross validation results
Parameter
Inside 90 Percent Interval 90.740
Inside 95 Percent Interval 95.061
Mean Error -0.543
Root-Mean-Square (RMS) Error 29.248
Mean Standardized Error -0.007
RMS Standardized Error 0.956
Root Mean Square Standardized Error: This value should be close to 1
Results - Empirical Bayesian Kriging Regression
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Results - Map of the GWR Rn estimates
•Simple Kriging have been applied to map the estimated values by GWR. Parameter
Mean Error 0.200
RootMeanSquare (RMS) 29.84
RMS Standardised 0.900
Average Standard Error 29.02
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
ResultsGeogenic Radon Potential
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Results – Radon Indoor Survey
Caprarola N AM (95% CI) GM(95% CI) Median Min Max Std.Dev.Indoor Rn
Bq/m3 82 677(491-863) 405(320-512) 364 14 5800 835
Preliminary and short-time indoor radon measurement surveys were
carried out in the Municipality of Caprarola by using charcoal
canisters, in order to select buildings in which integrated seasonal
radon measurements (in winter and summer) will be conducted
About 65% of the measured buindings are above
the limit indicated in European Directive 2013/59/EURATOM
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
GRP vs Indoor Rn
Results (preliminary indoor radon survey)
TGDR vs Indoor Rn
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
1 = center of the room, on the floor
2 = center of the room, 1 mt above the floor
3 = 1.5 mt from the walls
4 = on the walls
Results – Indoor Gamma Dose Rate
Caprarola N AM (95% CI) GM (95% CI) Median Min Max Std.Dev.
IGDR
mS/h116 0.48 (0.45-0.51) 0.46(0.43-0.49) 0.49 0.18 0.84 0.16
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
IGDR vs Indoor Rn
The graph shows a good
relationship between indoor
gamma dose rate and indoor
radon concentrations, thus
confirming the contribution of
the building materials for the
Caprarola site.
Results (preliminary indoor radon survey)
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
The application of multivariable spatial techniques seems to be more appropriate to
construct maps of the radon potential. The EBKR model has proven to be the best
spatial regression model among the other classical (OLS) and spatial (GWR) model
estimation.
GRP values estimated in correspondence of IR mesurements at ground floors and
cellars do not show a clear correlation with the IRC. However, IRC above 600Bq/m3
show a positive correlation with the terrestrial gamma dose values thus suggesting
that in this case IRC could be mainly due to the radon emanation from the building
materials (mainly tuff) rather than the geogenic potential of the soil and outcropping
rocks.
The good relationship between indoor gamma dose rate and IRC confirms the
contribution of the building materials for the Caprarola site.
Conclusions
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
THANK YOU FOR
THE ATTENTION!!
www.liferespire.it and www.liferespire.eu
14th INTERNATIONAL WORKSHOP – GARRM, September 18th – 20th(21nd), 2018, Prague, Czech Republic
Principal Component Analysis Results
Factor Loadings (Varimax raw) (Rn_SpatialRegression) Extraction: Principal components (Marked loadings are >.700000)
Factor - 1 Factor - 2 Factor - 3
40-K 0.738 -0.292 0.337
EmanationCoeff -0.762 0.277 -0.032
Pemeability -0.168 -0.503 -0.259
226-Ra 0.273 -0.194 0.794
Rn in the Pore 0.036 0.144 0.859
232-Th 0.778 0.218 0.176
220-Rn -0.054 0.873 0.085
238-U 0.638 0.500 -0.020
DTM 0.105 -0.760 0.216
Cumulative
Expl.Var (%)30.3 56.7 66.8